支持向量机在杂草分类中的应用
首发时间:2009-06-25
摘要:本文利用计算机图像处理技术和支持向量机识别方法研究了杂草的分类。首先对采集到的杂草叶片彩色图像进行预处理和分割,提取各类杂草叶片的形状、纹理特征参数。然后选取最有效的特征数据组合输入SVM进行交叉检验式的学习分类训练,实现杂草的有效分类。试验结果表明:使用该方法获得的杂草识别效率较高,不同分类核函数的识别效果不同,其中径向基核函数最适合杂草的分类识别。
For information in English, please click here
The application of support vector machine in weed classification
Abstract: This paper presents a weed recognition method with the combination of computer image processing technology and support vector machine classification.At first, weeds are segmented from the images and their shape and texture parameters are extracted.Then,the most effective combination of feature datas are selected to import to SVM to study classification training and the weeds are identified effectively. Experimental results show that the weed recoginition rate of the proposed method is high.Among three different types of kernel functions,radial basis function is the most appropriate for weed classification.
Keywords: weed classification support vector machine shape character texture character
基金:
论文图表:
引用
No.3342048260512459****
同行评议
共计0人参与
勘误表
支持向量机在杂草分类中的应用
评论
全部评论0/1000